There are three broad paradigms concerning the relationship between science and religion. In the first paradigm, science and religion are in conflict; in the second, they independent; in the third, they are to be in dialogue, or even joined together. I primarily advocate the independence paradigm, but I disagree with the usual formulation of it. The usual formulation to which I refer is the idea of NOMA (non-overlapping magisteria). NOMA is a concept invented by Stephen Jay Gould that states that science and religion cover two mutually independent realms. More specifically, science occupies the empirical realm of fact and theory, while religion covers ultimate meaning and moral values. The trouble with NOMA, in my view, is that it simplifies both science and religion.
Science simultaneously covers more and less than Gould seemed to think. Science covers the empirical realm, yes, but in principle, it can cover nearly anything ever thought. Science can consider any hypothesis--absolutely any--and ask, "What are the consequences of this being true?" Either there are some consequences, or there are not. If there are consequences, you can, in principle, test the hypothesis. If there are no consequences, well, it's hard to imagine that there are many hypotheses that are important, but have no consequences.
Of course, in practice, science doesn't cover quite so much. In practice, many hypotheses have important consequences, but are too difficult or not worthwhile to test. Furthermore, science in practice, is quite a bit more rigorous than the above description would suggest. You can't really disconnect the scientific process from the scientific establishment and still call it science. If there are no peer-reviewed journal articles on the topic, it's not really science. Maybe, in principle, it's scientific, but it has none of the authority we would accord to what we would normally call science.
As for religion, NOMA also gets its realm wrong. Religion is about far more than just meaning, purpose, and morality. Religious people are wrong if they think religion is all about believing this and believing that. Religion is also about ritual, community, devotion, and the general human experience. Often times, beliefs about ultimate reality are the ground for these practices, but they're not really necessary. You could hypothetically choose to interpret the beliefs semi-symbolically, or try a religion that has completely different beliefs, yet fulfills the same needs.
Of course, many religious people emphasize right belief, and it is not my place, from the outside, to say that this is any more or less true to religion. I can't just look at fundamentalists, who take a literal interpretation of the Bible to be essential to their faith, and declare that they're not doing religion right. But ask any religious person which is more fundamental: believing in miracles, or loving God? The latter is not in itself a belief.
So perhaps my reader has already begun to see the argument I intend to make. Science and religion are independent because, first, scientific practice simply doesn't cover most of what we consider religious beliefs, and second, religious beliefs are not the only component of religion. In a sense, they are NOMA after all, but Gould wrongly emphasized certain aspects just so he could fit them into neat little parallel boxes.
That's not to say you can't criticize religion. Obviously, I think you can. But the sort of criticisms that you can level against religion are not scientific. Maybe they're scientific "in principle", but no matter how you move around the definitions, it will never have the authority of science. Disagreeing with scientists about evolution is plainly ridiculous; disagreeing with scientists on religion is no worse than disagreeing with pundits on politics, or philosophers on philosophy. In fact, I would say the criticisms against religion belong in the category of philosophy. Some even in the philosophy of science. But that's not science.